Y.U.M.I - Yelp User's Maturity Index

Alessio Benedetti
October 27th, 2015

This work is licensed under MIT License

Highlight Yelp Users

The objective is to use the Yelp data to create an index the “Y.U.M.I”: Yelp User's Maturity Index that highlights “mature” users.

High rated YUMI users would give a way to easily choose between two or more businesses.

Methodology

The approach followed was to:

select relevant dataframes and fields
establish the metrics
build a predictive model to calculate the YUMI of an user
clusterize the YUMI to select high rated users (Gold, Silver & Bronze users)

Results 1/2

From the metrics,

\[ \begin{aligned} M_{1i}=\frac{WS^r_{i}}{WS^r_{max}} \qquad M_{2i}=\frac{WS^y_{i}}{WS^y_{max}} \qquad M_{3i}=\frac{friends_{i}}{friends_{max}} \end{aligned} \]

we built the following YUMI for the i-eth user:

\[ \mathit{YUMI}_i^p=\sum_{k=1}^3\alpha_k{M}_{ki}\qquad where \qquad \sum_{k=1}^3\alpha_k=1 \ and \ 0 \leq \alpha_k \leq 1 \]

Results 2/2

Sample variations

From the study we observed that in a high population, we may have a low probability to encounter high rated YUMI users (Gold, Silver & Bronze users).